TREASURY DEPARTMENT UNITED STATES PUBLIC HEALTH SERVICE HUGH S. GUMMING, Surgeon General STUDIES OF RECONSTRUCTED MILK BY ALBERT F. STEVENSON Sanitary Engineer GEORGE C. PECK Scientific Assistant AND C. P. RHYNUS Assistant Sanitary Engineer United States Public Health Service -(;-z;^'l1 REPRINT No. 608 FROM THE PUBLIC HEALTH REPORTS August 27, 1920 (Pages 2011-2045) WASHINGTON GOVERNMENT PRINTING OFFICE 1920 ADDITIONAL COPIES OF THIS PUBLICATION MAY BE PROCURED FROM THE SUPERINTF.NDENT OF DOCUMENTS GOVERNMENT PRINTING OFFICE WASHINGTON, D. C. AT 5 CENTS PER COPY V f^. Af De DEC 11 1920 t4 STUDIES OF RECONSTRUCTED MILK.* By Albert F. Stevenson, Sanitary Engineer; George C. Peck, Scientific Assistant; and C. P. Rhynus, Assistant Sanitary Engineer, United States Public Health Service. FOREWORD. By Earle B. Phelps, Consultant, United States Public Health Service. Pure milk contains a variety of substances, which, for convenience in analysis and reference, are conventionally divided into three groups : Fats, solids not fat, and water. The proportions of these three groups in the mixture vary with the breed of cattle, the feed, the season, and other variable conditions, but it will be sufficiently accurate for the present purposes and for illustration to say that normal milk contains 3J per cent fats, 9 per cent solids not fat, and 87J per cent water. At a time when, in the interests of food conservation, the possibili- ties of desiccated foodstuffs are being developed and their economic advantages exploited, it is unnecessary to dwell at length upon the purely economic benefits which would accrue were it possible to deal with the milk supply of the country and, especially, with its problems of storage and transportation, upon a water-free basis. Desiccation in general has two principal advantages: It reduces the weight of the commodity, thereby simplifying the problem of distri- bution ; and it improves the handling and keeping qualities of perish- able foods by doing away with the watery environment necessary for microbial activity. These advantages make it possible to distribute foodstuffs over wide geographical areas and thus to provide markets for those regions in which production is deficient or lacking. They also make possible a seasonal distribution or leveling, whereby the commodity may be carried over from a season of excessive production into a season of nonproduction. To these manifest advantages, which are common in the entire field of desiccated food, there is added, in the case of milk supplies, a third of equal if not greater importance. This has reference to the sani- tary aspects of the milk problem, which have so occupied the atten- tion of public-health workers in recent years. Milk provides so favor- able a medium for the growth and multiplication not only of normal and relatively harmless bacteria but also of the specific organisms of disease that the entire development of this important food supply, '■ Made under the supervision of Earle B. Phelps, Consultant, United States Public Health Service. Reprint from the Public Health Reports, vol. 35, No. 35, Aug. 27, 1920, pp. 20U-2045. 3 4 STUDIES OF RECONSTRUCTED MILK. especially as it relates to large cities, has been impeded and very seriously complicated by the necessary public-health restrictions. The storage and shipping of milk must be done at low temperature and can not extend beyond a few days at most, and pasteurization has come to be a recognized essential feature in the handling of a city milk supply. If, now, it were possible to desiccate the product in its fresh con- dition by a process which would greatly reduce the existing bacterial life and prevent the further multiplication of the surviving organisms, the limitations referred to would in large measure disappear. The actual realization of these conditions and the possibility of supplying to a community a milk reconstructed from its previously desiccated components and water, have been made possible by a series of steps, the commercial development of which has been taking place for many years. These are the mechanical separation of the fats in the form of cream from the skimmed milk^ and modern proc- esses of butter making; the development of processes, first, of skunmed milk evaporation, and, finally, of complete desiccation to a powder containing, in readily soluble form, all the milk solids except the fats; the development of mechanical means of emulsi- fying butter fat in water or skimmed milk solution, thus reconstruct- ing the cream; and, finally, the conception of reconstructing a whole milk, by emulsifying butter fat into a solution of milk solids of proper strength. This paper is a report of what is believed to be the first experiment upon a commercial scale with reconstructed milk. While the advan- tages of the plan, under many conditions, are so obvious that the idea had frequently been discussed, there were at the outset of this investigation certain difficulties to be overcome. In order to prevent fraud and protect health, it has been necessary for the States and the Federal Government to enact food laws which prohibit, in large measure, modification of natural food products. This raised the question of the propriety of the proposed procedure, which admit- tedly makes fraud, in the terms of the pure food law, easy and its detection difficult. The particular situation which has been created by the possibilities of reconstructed milk is without precedent, and, as it deals with one of the most fundamental of the human foods, a most conservative attitude upon the part of the officials charged with the enforcement of the pure food laws has very properly been maintained. In view, however, of the tremendous advantages which are foreseen in the development of this field, it is believed that the problem must be handled entirely upon its own merits, and with the utmost frankness. There can be no question of the impropriety of handling reconstructed STUDIES OF RECONSTRUCTED MILK. 6 milk, except under its own label and with complete information as to its source and method of preparation. If it can be shown, however, that reconstructed milk can be marketed in a community under cleaner and safer conditions and at less cost than can ordinary milk, the consumer is entitled to these advantages just as he is also entitled to know exactly what he is purchasing, and, if he prefers to use fresh cows' milk, to be assured that he receives that for which he asks and pays. A second possibility which has prevented more active development of reconstructed milk has been the fear that the public might be slow to appreciate its advantages and reluctant to give up a part of the dietary so well established as milk in favor of what must appear to be a sort of manufactured product. While this argument will carry no weight whatever among physicians, dietitians, and others competent to look into the subject with sufficient thorough- ness, it is a matter to be reckoned with most thoughtfully by one who would undertake the commercial development of a market for reconstructed milk. In the situation which arose at the new Government city of Nitro, W. Va., both of these objections lost most of their force. A city to house some 25,000 people was being built overnight, and w^as without any visible milk supply, nor did it seem probable that a sufficient supply of safe milk could be obtained. The health admin- istration of this city was to be in charge of the United States Public Health Service, so that there could be no question of conflict with traditional health regulations. The conditions made it unnecessary to fear competition with normal milk, so that it was practically assured in the beginning that, if the milk was satisfactory, a market would be secured. Under these favorable conditions a plan which had been gradually maturing was put into effect, and the present paper is a report of the results obtained. Unfortunately for this particular purpose, the experiment had to be discontuiued before the complete data that were hoped for had been gathered. It is believed, however, that the results obtained are of so great impor- tance that the outcome of this first commercial demonstration of the possibilities of reconstructed milk should be recorded. Mr. Stevenson designed the plant and prepared the specifications for the mechanical equipment and has been wholly responsible for the planning and execution of the experimental work. He has been ably assisted throughout by Mr. Peck who was in resident charge during the construction period, and who, in the capacity of super- intendent, was directly responsible for the operation of the station. Mr. C. P. Rhynus made the bacteriological studies, and Mr. Leslie Z. Peck served as assistant superintendent. 6 STUDIES OF RECONSTRUCTED MILK. I. THE MANUFACTURE AND HANDLING OF RECONSTRUCTED MILK AT NITRO, W. VA. By Albert F. Stevenson, Sanitary Engineer, and George C. Peck, Scientific Assistant, United States Public Health Service. Introduction. Nitro, W. Va., is located on the Kanawha Eiver, about 13 miles from Charleston. It was built by the United States Government to house the laborers and mechanics employed in building the United States Explosives Plant C and later to house persons permanently employed in the smokeless-powder plant. Accommodations were made for about 25,000 inhabitants. The several institutions and industries necessary to form a well-organized conmiunity, such as schools, churches, hospitals, department stores, restaurants, water supply, sewerage, abundant food supply, etc., were provided. In short, a modern city was constructed on the land surrounding the explosives plant. The construction of this city and the organization of the various necessary secondary industries was accompanied by many difficulties which were increased by war-time conditions. The gathering together of sufficient food of satisfactory quality proved to be one of the serious difficulties, and radical departures from current practice were made in order to overcome them. One of the most interesting and instructive divergencies was made in securing a sufficient quantity of fluid milk to satisfy the demand. The present article deals with this part of the work. Available Normal Source of Milk. Very little milk is produced in West Virginia in the vicinity of Nitro. The region is mountainous and little if any natural pasturage is available. In fact, previous to the building of the various Govern- ment industries located in this section, hardly enough local milk was available to supply the normal Charleston market. Up to the time Nitro was conceived, little attempt was made to increase the supply. Some milk was shipped in from the dairy section of Ohio, but with the advent of war, the scarcity of milk and the congestion of railroad traffic made this source undependable. The United States Public Health Service was called on to suggest some method of procuring a milk supply for Nitro. Reconstructed Milk Products. For pome time it has been known that a liquid closely resembling milk and cream could be made by emulsifying butter fat obtamed from unsalted butter in a solution of skimmed milk powder or diluted evaporated skimmed milk. This procedure has been very widely STUDIES OF RECONSTRUCTED MILK. 7 used by the ice-cream industry and has been sanctioned for this piu"- pose by the -pure-food officials. This liquid has also been made in small quantities at some Army field hospitals, and on several of the battleships to furnish a supply to the officers' mess. It has been made at various dairy and milk shows as a means of advertising milk powders. The manufacture of reconstructed products on a small scale led to the assumption that a fairly large-scale plant could be successfuUy operated and the milk supply of a city the size of Nitro manufac- tured. After conferring with officials of the department of health and sanitation and the commissary department at Nitro, it was decided to recommend the building and equipping of a plant of suffi- cient capacity to supply the entire city with reconstructed milk and cream. It was certain that milk could be made which would be satisfactory if consumed within a few hours of the time of manufac- ture; but little, if anything, was known about the cost of manu- facture, the method of handling, and the keeping qualities of the products. In order to make a thorough investigation of the subject and at the same time supply Nitro with milk, the Public Health Service undertook to design the plant and superintend its operation as long as the process was in the experimental stage. An equipment was de- signed which, it was estimated, would handle 2,000 gallons of bottled milk in an eight-hour day, and which capacity could be increased to 3,000 gallons per day by the addition of another pasteurizing and emulsifying unit. Milk and cream could be handled in either bulk or bottled form, although equipment of sufficient size to bottle all the output was provided. It was expected that all milk sold, except that used for cooking, would be served in bottles. Description of Building. The "milk plant" was located in the south end of the reserva- tion in a one-story frame building 150 feet long and 51 feet wide. This same building also housed a small cold-storage warehouse used by the commissary department for the storage of perishable foods. The space allotted to the manufacture of reconstructed milk was 123 feet long and 51 feet wide. This space was somewhat excessive, but on account of the experimental character of the work, extra room was provided. The general plan of the building is shown in Figures 1 and 2, and a sectional elevation in Figure 3. The building was divided into eight rooms: A milk manufacturing room; a room for the washing of containers; cold-storage space for milk and butter; a dressing room; a storage room for milk powder; an office; and a room for the refrigerating machinery. STUDIES OF RECONSTRUCTED MILK. 4 1 , >« < !• owders made by other processes, and the purchaser must decide on the product best suited to his needs. In the manufacture of reconstructed milk, the follow^ing points are of extreme importance in the selection of a skimmed-milk powder: The powder should be wholly and easily soluble in water and when reconstructed should give a solution with the characteristic flavor of normal skimmed milk. The various constituents, such as the sugar, proteins, and mineral salts, should not have been altered during drying, and upon recon- struction should appear in as near the original state as possi])le. A solution of the powder should have the powder to hold an emulsion of butter fat similar to that found in normal milk. As the reconstructed milk must meet all the legal requirements of normal market milk, the original milk used for drying must be pro- duced in a cleanly fashion, and must meet all requirements of the local health department of the district where the reconstructed milk is to be sold. The process of drying should also be conducted in a cleanly manner, and rules prescribed for the handling of normal milk should be enforced in the manufacture of the skimmed-milk powder. It is needless to say that this powder should contain only and all the ingredients occurring in normal skimmed milk, with the exception of the water. Skimmed-milk powder is deliquescent and absorbs moisture rapidly. If the moisture content is high, the powder will lump in the containers and be unfit for use. It should therefore be stored in moisture-proof containers and should be sold with a guaranteed maximum moisture content. Butter fat. — Sw^eet or unsalted butter is used as a source of butter fat. A good quality of sweet butter is almost as difficult to obtain as a good quality of skimmed milk powder or water. It should have good flavor, be free from pathogenic organisms, and be manu- factured in such a manner that objectionable flavors and odors will not be produced in it during storage. It should never be artificially colo»"ed, and manufacturers should guarantee the butter-fat content. The butter should be shipped in proper containers, and every con- tainer should be scored by the purchaser before it is accepted. Butter should be stored at a temperature sufficiently low to prevent the production of disagreeable flavors and odors and should be taken out of storage only as needed. Commercial Manufacture of Reconstructed Milk. RECONSTRUCTION OF SKIMMED MILK. If the skimmed milk powder to be used has been selected with proper care the process of dissolving it is not a difficult one. At KiUo a val buUermilk machine of ordinary design was used for this 16 STUDIES OF RECONSTRUCTED MILK. purpose. This machine was of 300 gallons capacity and was equipped wdth a revolving heating coil to which blades were attached, serving simultaneously as agitator and heater. In selecting a vat, care should be taken to obtain one with a cylindrical bottom, and sides which are tangential to the surface of the cylinder. There should be no corners in which the undissolved powder may be pocketed. The necessary amount of distilled water should be placed in the vat and a weighed amount of skimmed milk powder added. The powder may be dumped in directly from the barrel, if care is taken to prevent foreign matter from falling from the sides of the barrel during the process. It was found that better results were obtained when the powder was removed from the barrel with a large sugar scoop. The coil agitator should be in operation while the powder is being added. The temperature of the water during the period of solution of the powder should be between 70° and 80° F. Adding butter, and pasteurization.— After the complete solution of the powder had taken place, the skimmed milk was pumped to one of the mixing vats on the balcony. Here the necessary amount of butter, which had previously been cut into 4-inch cubes on the porcelain topped table provided for the purpose, was added, and the mixture brought to a temperature of 146° F., and held there for 30 minutes. It was found that by the time the temperature had reached 146° F. all the butter had melted. The holding time for the pasteuri- zation of the skimmed-milk-butter mixture may therefore begin when the correct temperature is reached. It is necessary, however, to cut the butter into 4-inch cubes or less if this procedure is to be followed. Emulsifying the hutterfat. — After the mixture had been pasteurized, and while it was at the pasteurizing temperature, it was passed through the centrifugal emulsors. Emulsification is brought about by the forcing of the mixture of butter and skimmed irdlk through an extremely narrow opening, using centrifugal force generated by revolving the bowl of the emulsor at a speed of approximately 15,000 revolutions per minute. Cooling. — From the emulsors the hot reconstructed milk was con- ducted to the upper trough of the cooler through sections of 1^-inch sanitary milk pipe to which had been attached small conductor heads. The milk from the emulsors contained a great deal of foam, and difficulty was experienced in transmitting this foaming product to the cooler. A large bank of foam collected where the milk entered the conductor head and also in the upper trough of the cooler. This necessitated the providing of ample capacity at both of these points to prevent the foam from accumulating and running over the edge of the conductor head and trough. For this reason it is felt that the use of an external cooler should be avoided in this process. The use STUDIES OF RECONSTRUCTED MILK. 17 of an external tubular cooler would eliminate to a great extent this waste and .inconvenience. By inserting an antifoam tank, as described below, between the emulsor and the cooler, much of this trouble might be eliminated. This arrangement was not tried at Nitro, however. A proper regulation of the temperature of the milk as it leaves the cooler is an important point in the process of manufacture. As with ordinary milk, freezing throws the butterfat out of emulsion to a greater or less degreee. The flow of brine should therefore be regu- lated so that a minimum amount of milk freezes to the cooler. All milk which does so freeze should be re-emulsified before bottling. Satisfactory results can be ob- ' ^' — .f2f^M Level ^'V* .Milk /^Ourter Milk Level 16". Milk iENTERS; - 1 ^ Fig. G. — Section of aniifoam tank used at the Recon- structed Milk Plant, Nitro, W. Va tained if the milk leaves the cooler at a temperature of from 40° to 50° F. Foam removal. — As has been stated, the reconstructed milk as it leaves the emulsor con- tains a great deal of suspended air. This air is in such a finely divided state that it is elim- inated slowly. Bottles filled with the milk directly from the cooler show, on two hours standing, an air space between the cap and the surface of the milk varying from 1 inch to 3 inches in height. In order to eliminate this foam an antifoam tank was inserted between the cooler and the can and bottle fillers. This foam tank is simply a detention tank, from the bottom of which the milk is drawn. A sketch of it is shown in Figure 6. With the emulsors running at full capacity (approximately 400 gallons per hour) this tank gave a detention time of approximately Z\ minutes. This arrangement eliminated practical^ all tlie foam from the milk. Filling containers. — From the foam tank the milk passed into a rotary bottle filler, A "no foam" can filler was inserted between the antifoam tank and the bottler. At this point all the cans were filled. The bottle filler was driven by a direct-connected motor and had a capacity sufficient to handle the effluent from the emulsors. This type of filler and capper has a decided advantage over a case filler in that all bottles have to be handled twice and therefore get two inspections. When using the case filler there is a great tendency to slight the bottle inspection, and many damaged bottles which 12325°— 20 3 18 STUDIES OF RECONSTRUCTED MILK. should not be used are filled and put into circulation. The bottles used were of the ordinary glass, common-sense type, capped with paper caps. As is customary, the caps bore the package label. Storage. — Immediately after the containers were tilled they were trucked into the cold-storage room, where they remained until delivered. The temperature of this room was approximately 33° F. Labeling. — ^Regulations pertaining to all foodstuffs should be ap- plied to reconstructed milk products. These products should be sold only for what they are and only when the packages are properly labeled. According to the present law, a milk made from skimmed milk powder, water, and butter is not "milk," and should not be sold as such. The law distinctly states that milk is the normal fluid secreted by the mammary glands of the cow, and a product which has been reconstructed certainly can not be classed as "milk." In choosing a name for this product various titles have been con- sidered. From the list has been chosen the term "reconstructed milk." It is felt that this is fully descriptive and leaves no reason Fig. 7. — Bottle caps used at the Reconstructed Milk Plant, Nitro, W. Va. for doubt as to the product. The word "reconstructed" may pos- sibly arouse some sense of fear in the minds of the unenlightened con- sumer, for it may signify to him a product which has been at some time in an unsatisfactoiy condition and later renovated. Various other names which have been suggested are "remade," "emulsified," "reconstituted," "recombined," and "rehydrated." Any of these terms may be used, subject to the approval of the officials who are intrusted with the enforcement of the pure-food laws. In labeling any of these manufactured products the word "recon- structed," or any other similar word which may be used, should be given as much prominence as the word "milk." The label should also set forth the percentage of butter fat and the percentage of milk solids not fat which the product contains. Figure 7 shows samples of the various caps used at Nitro. The reconstructed milk manufac- tured there contained 9 per cent milk solids not fat and 3h per cent butter fat. The cream contained 25 per cent butter fat. A form of fermented milk similar to buttermilk was manufactured and styled STUDIES OF RECONSTRUCTED MILK. 19 "Nitrolac." It was not considered proper to label this milk butter- milk, for the term "buttermilk" has a distinct meaning. It is a product of the churning of milk or cream and the name should be used for no other product. The fermented milk used at Nitro was simply a cultured milk of a low percentage butter fat. Washing containers. — ^All containers were washed in the room pro- vided for that purpose. The bottles were inverted and washed in the cases. The washer used consisted of a series of tanks from which water or alkali solutions were pumped with considerable pressure through a series of jets. The cases were intermittently advanced by a mechanical device and the jets so arranged that during the resting period a powerful stream of water or chemical solution was injected into each bottle while a number of streams were forced against the outside and inside of the case. The temperature of the first rinse water was approximately 110° F., a temperature sufficient to warm the cold bottle without danger of breaking it. The temperatures of the subsequent washing and rinsing solutions were so regulated that the final temperature of the bottle was approximately that of boiling water. Before leaving the machine each bottle was subjected to a jet of live steam which acted as the final sterilizing agent. All cans and metal containers were washed by hand in a galvanized iron tub provided for the purpose. After a thorough brushing with an alkali sc^lution, the can and cover were inverted over a can rinser and sterilizer where they received a hot-water rinse and a thorough steaming. Cleaning milk-handling machinery. — At the end of each day's run all milk-handling machinery and piping were completely disassembled, rinsed with warm water, scrubbed with a brush and hot alkali solu- tion, rinsed with hot water, and then sterilized by blowing live steam into or through them for a considerable time. The machinery was then partially assembled. In the morning before operations com- menced, the machinery was completely assembled and connected at various points to the steam line. Steam was blown through the assembled machinery for approximately 20 minutes. This produced a satisfactory sterilization. The most difficult piece of machinery to sterilize was the tubular cooler. This is the general experience where this type of cooler is used. At the end of the run the tubes are full of water or brine, and the ordinary methods of steaming are not sufficient to heat the whole mass up to the sterilizing point. Satisfactory sterilization may be accomplished by emptying the tubes and either blowing steam, under pressure, through them, or by inserting a steam hose between the covers and the cooler and allowing steam to blow in for some time. If either of these methods is used, a by-pass around the brine pump must be supplied so as to drain the brine coil back to the tank without 20 STUDIES OF EECONSTRUCTED MILK. wasting the brine. These methods of sterihzation are very likely to cause leaks in the cooler, ov/ing to the uneven expansion of the coils. It was found at Nitro that satisfactory sterilization could be accom- plished by a thorough brushing with soda solution and then rinsing with approximately 200 gallons of water which had been previously heated to about 180° F. in one of the pasteurizing vats. From a bacteriological standpoint this method is not as efficient as draining the coils and appl3ang steam. m QUANTITY OF THE RECONSTRUCTED PRODUCTS MANUFACTURED. As has been stated, the plant had a capacity of at least 2,000 gallons of reconstructed products in an eight-hour day. This amount of milk was never sold in Nitro, owing to the impossibility of proper delivery. Nitro was never completely finished, and the streets, up to the cessation of operations, were in such poor condition that a house-to-house delivery could not be made. It was intended to sub- stitute for this retail delivery a wholesale delivery to a number of small distributing stations centrally located, where the various milk products could be purchased by the consumer and carried to the homes. This system would have had the advantage of elimJnating bottle loss. Unfortunately, these milk stations were not finished in time to take care of the rush. The fact that milk could be purchased only at the general stores cut the consumption greatly. Table I gives the daily quantities of the various products manu- factured during the months of September, October, November, and December, 1918, and part of January, 1919. Table I. — Quantity of reconstructed milk products manufactured daily at Nitro, W. Va,, during Sep te^nber, October, November, and December, 1918, and part of January, 1919. Date. Pounds of milk. Pounds of cream (2.5 per cent). Pounds of ice cream mix. Gallons of ice cream. 1918. Sept. 6 1,091 9 1,703 10 2,200 11 3, 000 12 3,500 13 . 3, 400 14 4,400 15 2,200 16 0,400 17 5,400 18 . ... . 4, 000 19 7,200 20 . . 4,800 21 . 2,400 22 4,400 23 4,800 24 . . . . 7, 200 25 4, 800 26 7, 200 27 7, 200 28 0, 3 ) ) 29 7, 2.M) 30 9, liOO 500 s 10 1,000 10 10 ::::;.:::.! ;... 10 10 1,200 25 STUDIES OF KECONSTRUCTED MILK. 21 Table I. — Quantity of reconstructed milk products manufactured daily at Nitro, W. Va., during September, October, November, and December, 1918, and part of January , 1919 — Continued. Oct. Nov. Dec. 11.. 12.. 13.. 16.. 17.. 18. 19.. 20.. 21.. 22. 23. 24. 25. 26. 27. 28. 29. 30. 1. 2. 3. 4. 5. 6. 7. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Date. 1918. Pound.s of milk. 7,20!) Pounds of cream (25 per cent). 7,2;X) 7,200 7,200 7, 200 7,200 7,200 7,200 7,200 7,200 7,200 7,200 7,210 7,200 7,200 7,200 8,400 7,200 7,200 8,400 8, 100 8,400 8,400 9,600 8,400 8,400 9,000 9,600 7,200 8,400 7,200 7,200 8,400 8,400 7,200 8,400 8,400 7,200 12,000 9,600 8,400 7,200 12,000 4,800 4,SO0 7,200 4,800 4,800 7, 200 4,800 4,800 2,400 4,800 4,800 4,800 4,800 4,800 4,800 2,400 4,800 2,400 2,400 4,8(X) 2,400 7, 200 2, 400 2,400 4, 800 4,800 4,800 2, 400 2, 400 2,400 4,800 500 500 Pounds of ioo cream mix. 500 500 500 500 500 500 1,200 1,210 1,200 1,200 1,200 Gallons of ice cream. 1,200 1,200 600 600 1,200 1,200 1,200 1,200 1,200 1,200 1,200 1,200 1,200 22 STUDIES OF RECONSTRUCTED MILK. Table I. — Quantity of reconstructed milk products manufactured daily at Nitro, W. Va.y during September, October, November, and December, 1918, and part of January, 1919 — Continued. Date. Pounds of milk. Pounds of cream (2.5 per cent). Pounds of ice cream mix. Gallons of ice cream. Dec. 20 1918. 4,800 2,400 2,400 2,400 2,400 2,400 2,400 2, 400 2,400 2,400 2,400 2,400 60 21 45 23 500 80 24 1,200 00 23 . 45 27 . 500 50 28 45 30 1, 200 55 31 - 30 Jan. 1 . ... 1919. 1,000 CO 3 600 60 4 500 5 35 6 2,400 2,400 2,400 30 7 : 45 8 1,200 80 9 50 10 2,400 4,800 500 45 11 40 12 65 13 500 1,200 80 i; 2,400 2,400 CO 15 85 If) 1 40 1 The quantities of milk and cream produced are large enough to serve as a basis for economic considerations. OPERATING SCHEDULE. It was expected that full data on the cost of operation of this plant could be collected by keeping accurate labor charts and meter- ing the electric power, the water, and the steam consumed by each machine. Unfortunately the meters were lost in transit, due to congested traffic conditions, and this part of the data is not available. Time sheets of the various operations were kept, and the operating schedule given below was computed from them. The capacity of the plant depends on the time required to mix the ingredients and pasteurize them. It is therefore of first import- ance to determine the time required for the various operations con- nected with this part of the process. Accurate time sheets were kept for two months, and it was found that the following time inter- vals were required to perform the various operations included in the mixing, pasteurizing, and emulsifying of one batch, 282 gallons, of milk: Minutes. Weighing and transferring water and skimmed milk powder to vat 33 Dissolving powder 10 Emptjing skimmed milk vat IS Preparing and adding butter 13 Heating mixture to 146° F 28 Holding for pasteurization 30 Emulsifying (2 emulsors in use) 45 STUDIES OF RECONSTRUCTED MILK. 23 Using these data, a daily schedule of operation of the three vats and two emulsors was planned. This schedule is given in Table II. Vats A and B represent the mixing and pasteurizing vats on the balcony, and vat S is the sldmmed milk vat located on the main floor. It will be noted from Table II that vat S is in continuous opera- tion, but that vats A and B are idle 20 minutes, and the emulsors 15 minutes, between every two batches. This loss of time could be eliminated by changing the method of preparing the skimmed milk. Too much time is consumed in measuring and handling the water and powder. This time can be reduced by arranging the building so tha*^ the powder storage room and the distilled water storage tank are on the same level witli the present balcony. A weighed charge of pow- der could tlien be dumped directly into a hopper placed above the skimmed milk vat while the powder in the previous batch is dissolv- ing. If, at the same time, water from the elevated tank could be run directly into the mixing vat through a 4-inch conductor, the time of charging the vat could be reduced to approximately 10 minutes. An ordinary gauge glass attached to the end of the vat, extending the full height of the vat, could be calibrated accurately enough to pro- vide means of measuring the distilled water. By using this arrange- ment of measuring devices and by having the man responsible for the mixing of the skimmed milk commence work an hour ahead of the rest of the force, the output of the plant can be increased from six to eight batches per day. Such an arrangement is outlined in Table III. It will be noted that between Batches II and III, IV and V, and VI and VII there are 12-minute intervals during which no milk is passing through the emulsors. These delays are unavoidable when vats of this capacity are used. 24 STUDIES OF RECONSTEUCTED MILK. Us 2S - tJb Tito g >> a> o r7^ d o) t- r] q; CD P oj J^ ^ ^Zla 2 • - ^ jT s"* c £? |f>.a.s.s.«> CLp. « 03 t3 f^ a g C3 03' C3' d.^" ~ a*^ +m aa "^OO O'-^CSC'l '•B ^ §00 £33 mO.03 Co" © O o3 a M'M p<> .S H 03; K Q- X! o -a 1 5 1 o o 03 C g > o ca a >-• — Mm Qj v.' G a^s + aa a cj 0-: -coco >;, w -g ■§ l>> a-s a a Oi a 5a a a ^assis o3 tS a"+aaa' o3 H c3 c3 c3 Oi ■ CO CO 05 ci^ rt fi D--^ S S S 8 S .a.' , aai a 03 +aaa+ ai ,^aag_-g-gg *" 03 a!;aaa + agg' ft ft a. g X o^x 00 "j tes, preparing tes, heating to tes, holding at tes, emulsifyin completed. utes, preparing i utes, heating to utes, holding at utes, emulsifyin [V completed. Batch VI. utes, preparing utes, heating to utes, holding at utes, emulsifyin VI completed. Batch VIII. es, preparing a es, heating to 1 es, holding at 1 es, emulsifying n completed. 3 3 3 ati 3 3 3 a 3 2 3 3> seIsI I'ias" Isssl 3 3 3 3'^ laEsI to^ OiO c3 tO^OiO^ COi-hOiOC3 + + + + - »-Hf— ICO'T'CQ ^-tr- es, preparing i es, heating to es, holding at es, emulsifyin Dmpleted. Batch III. es, preparing i as, heating to es, holding at es, emulsifyin completed. Batch V. es, preparing es, heating to es, holding at es, emulsifyin completed. ^3^^" minut minut minut minui tch VI 3 3 3 3'1 3 3 3 3> min min mill min tch I min min min min tch I min min min min tch CC ■-' O iC C3 to -< C O oi to — o ■' 5 + + + + . ++++ . + + + + . + + + + - a 9 a a a aasaa aaaaa saaaa (S 03 03 03 d caoJcOoJcS 03c3(jjnii (iddcid. O to 1^ I^ (M o e 0) measuring to dissolve emptying ^ Batch II. measuring dissolving emptying Batch III. measuring dissolving emptying Batch IV. , measuring dissolving emptying Batch V. measuring dissohing emptying Batch VI. measuring , dissolving , emptying Batch VII. 3>-5; ill minutes, minutes, minutes, empty. minutes, minutes, minutes, empty. minutes, minutes, minutes, empty. minutes, minutes, minutes, empty. minutes, minutes, minutes, empty. minutes, minutes, minutes, - empty. minutes minutes minutes ; empty. minutes, minutes, minutes, empty. +12] +10: + 16 1 , vat +12: + 10: + 16 , vat + 12 + 10 + 16 , vat +12 +10 + 16 , vat +12 + 10 + 16 , vat .+12 .+10 .+16 ., vat .+12 . + 10 . + 16 ., vat . + 12 . + 10 . + 16 ., vat aaaa aaaa saaa sdaa aaaa asaa asaa aasE c3 c3 03 C3 c3(SoJ(fl 03 030353 oc^ 5S8S S ■«• 'O t>^r>^odod cCGoooi^ o^o»aio> oooo rt' rt' ..; ..; ,j -4 c^i (M cj cj •-<■ ^' rH -i f-J (M 26 STUDIES or RECONSTRUCTED MILK. OPERATING FORCE. A plant as described can be operated with a force of seven men. This force should consist of a foreman, who would have general charge of the plant and would personally operate the skimmed milk vat; an assistant foreman, who would spend his entire time on the balcony level, operating the pasteurizers and emulsors; three men operating the bottling and capping machine — one man feeding the machine with empties, one man removing the filled bottles, and one man handling the cases of empty and filled bottles — one man trucking the filled cans to the refrigerator, and one man operating the refriger- ating machine and assisting as his time permits at any temporary point of congestion. ' The foreman and assistant foreman should be men skilled in the handling of milk and the operation of milk machin- ery, and the engineer should, of course, understand the operation of a refrigerating machine. The other employees may be of a good grade of laborers. COST OF PRODUCING BOTTLED RECONSTRUCTED MILK. The cost of producing bottled reconstructed milk will vary with the locality and with the market price of the various commodities used in the manufacture and handling of the product; therefore, no exact figures can be given. It can be stated, however, that the cost of a plant to manufacture reconstructed milk, and the number of men and the amount of power necessary to operate it, are no greater than similar items in the operation of an ordinary modern pasteurizing and bottling plant, and therefore, the cost of manufacturing and bottling reconstructed milk is approximately the same as the plant charge for pasteurizing and bottling fluid milk. The delivery charge is, of course, identical in both cases. The economic advantage gained by drying the skimmed milk and manufacturing butter at the point of production and reconstructing the milk at the point of con- sumption, rather than handling it in the normal fluid state, lies in the difference in price between sweet fluid milk delivered in bulk at the pasteurizing and bottling plant, and powder, butter, and water delivered at the same location, plus the economic advantage gained by always having the supply equal the demand. Skimmed milk powder at the present time is manufactured in the great dairy regions of New York, Pennsylvania, Ohio, Indiana, Illi- nois, Michigan, Wisconsin, Minnesota, and on the west coast. Sweet butter is also manufactured in quantity in these States. Distilled water can be produced locally. The prices of powder and butter are necessarily the market prices at the point of manufacture, plus the freight charges. During the past year (1919) the market price of skimmed milk powder at the factories has varied in carload lots from 20 cents to 26 cents per pound. Sweet butter has varied on the New York market from 46 cents to 71 cents per pound. Distilled water STUDIES OF RECONSTRUCTED MILK. 27 can be made for less than one-quarter of a cent per gallon. Using these figures, the cost of the ingredients entering into a quart of milk containing 9 per cent solids not fat and 3.5 per cent fat will vary from 7.9 cents to 13.3 cents. To this must be added the freight on the powder and butter, which is a minor item. The price paid the producers in the South, during this period, for normal fluid milk has varied from 10 cents to 20 cents per quart, depending on the locality. It is certain, therefore, that in these sections reconstructed milk can be sold for a lower price than normal milk. Whether reconstructed milk is more economical for a given com- munity than normal fluid milk can only be determined by a careful comparison of the prices of the raw commodities delivered at the point of consumption. If all the milk produced in the country could be dehydrated and subsequently reconstructed as needed to meet the demand, there is Uttle doubt that a great financial saving to the country at large would result. Until such method is universally adopted a careful study of the local conditions of each community must be made before deciding on the correct method of milk handling. From a bacterial standpoint reconstructed milk has decided advan- tages over market milk of good quality, as will be shown in the sec- tion on the bacteriology of this product. THE QUALITY OF THE FINAL PRODUCT. The manufacture of milk powder and reconstructed milk is in its infancy. The processes have been studied to some extent, but little is known regarding the effect of the various methods of drying and reconstructing on the composition of the finished product. Infant feeding experiments have been conducted in England, in which dried milk powder was used. These experiments are repoited in Food Reports No. 24 to the local government board on Public Health and Medical Subjects, New Series No. 116.^ These experi- ments indicate that children can be raised on reconstructed milk with no more difficult}^ than on normal cows' milk. To the casual adult drinker of milk, carefully produced recon- structed milk made from the better grade of powders available in quantity on the market to-day, has the appearance of normal milk, except that it lacks a cream line. It has a flavor slightly more ''cooked" than the pasteurized market milk. This flavor to many observers gives the impression of excessive "richness." It is not objectionable to those used to pasteurized milk, but is somewhat distasteful to those accustomed to drinking raw milk. The freshly made product very easily passes for nrst-cl'iss noiuial milk. One sc:ious physical defect exists, however: the fat emulsion as produced 'See also "Dried Milk Powder in Infant Feeding: Safety, Usefulness, and Comparative Value— A Preliminary Report," by W. H. Price, Surgeon (R.), United Slates Public Ileallh Service, Public Ilealtli Reports, vol. 35, No. 14, Apr. 2, 1920, pp. SU'J-ii2S.— Kuitoe. 28 STUDIES OF EECONSTRUCTED MILK. at Nitro is not wholly permanent. On standing for 48 hours at a temperature approximately 35° F. or for 20 hours at room tempera- ture, a thin crust of butter forms on the top of the fluid. If the milk is kept in an ordinary refrigerator and consumed within 24 hours no separation is noticed. This '' buttering " in all probability is caused by some change in the complex ingredients of the skimmed milk brought about by the drying process. The power to hold fat in emulsion has been partially destroyed. As soon as this effect is overcome by the manufacturers, a product will be available which will readily compete with normal fluid milk. At present, reconstructed milk forms an excellent emergency supply which may readily take the place of normal milk during a shortage of the latter product. Summary. 1 . Reconstructed milk and cream made from skimmed milk powder, unsalted butter, and water have been produced in large quantities at Nitro, W. Va., and sold to the public. 2. The cost of manufacturing these products in the Southern States is less than the cost of local normal milk. As the distance between the points of efficient production and consumption is less- ened, this diflFerence in cost becomes less. In the dairy sections at the present time fluid milk can be sold more cheaply than recon- structed milk. 3. Reconstructed milk products serve as excellent emergency supplies, and as soon as the process of manufacturing milk powder is perfected they wiU no doubt compete in the open market with normal milk products. Acknowledgments. The authors wish to express their appreciation of the valuable assistance given them throughout this demonstration by Passed Asst. Sm-g. J. A. Watkins, of the United States Public Health Service, resident officer in charge of the department of medical relief and sanitation of Nitro; Mr. Hugh C. Leighton, executive officer of the Nitro commissary department; Mr. C. S. Bassett; Mr. L. C. Johnson; Mr. R. G. Soule; and all others who through suggestions or material aid helped in making the demonstration a success. II. AN ANALYTICAL STUDY OF RECONSTRUCTED MILK. By Albert F. Stevenson, Sanitary Engineer, and C. P. Rhyntjs, Assistant Sanitary Engineer, United States Public Health Service. Before any process devised for the handling of milk can be pro- nounced a success, an analytical study of the product, both from a bacteriological and chemical standpoint, must be made. If this study shows the product to be materially altered or to be a potential source of danger, from a health standpoint, the process is valueless. It was therefore of the greatest importance to determine in this plant, the initial one producing reconstructed milk on a commercial STUDIES OF RECONSTRUCTED MILK. 29 scale, whether the product was of a high or low grade, from a bacterial and chemical standpoint. A plant laboratory was installed where bacteriological analyses could be made. It was possible by this means to check the plant operation and to determine, at all times, the quality of the product which was being sold. As is always the case, the field laboratory proved one of the most important assets of the plant. Samples of the finished product as well as samples from the various stages of the process of manufacture were analyzed daily. The finished product was again sampled and analyzed, after delivery, by the division of health and sanitation of the Nitro organization. These last analyses correspond to those which would be made by the board of health of an ordinary community. The methods employed in making bacteriological analyses were those recommended by the committee on standard methods of milk analysis of the American Public Health Association. Samples were collected from the various stages of the process of manufacture, in glass test tubes plugged with nonabsorbent cotton. Filled pint bottles, selected at different periods during the day, served as bottled milk samples. These samples were refrigerated until the actual analysis was made, which occurred, generally, within two hours of the time of sampling. The culture media used for the bacteriological analysis were made at the Hygienic Laboratory in Washington, D. C, and shipped to Nitro in tin containers. The media were transferred from the cans to 4-ounce glass bottles, after arriving at the plant, and were then resterilized. A check titration was always made on each ship- ment of media, after the final sterilization in the glass bottles. Sterili- zation of media was accomplished by heating for one hour at 100° C. in an Arnold sterilizer for three consecutive days, and then incubating the media for 24 hours to throw out any contaminated bottles. The dilution water was sterilized in the same manner. Dry sterilization was accomplished in the oven of an ordinary gas stove. Forty-eight- hour 37° C. total counts were made on the various samples. No attempt was made to isolate any particular organism. The results of the bacteriological determinations made on the finished reconstructed milk in the final containers are given in Table IV. Table V shows the results of arranging these counts in the order of magnitude. Seventy-five per cent of them were 5,000 or less. The highest count obtained was 55,000, and only three times was a count over 30,000 obtained. Total comits made on the bottled milk by the department of sanitation are given in Table VI. Rearrangement of these comits in the order of magnitude, as shown in Table VII, em})hasizes the fact that 15 per cent of them are imder 1,000 and 97 per cent are under 5,000. Only one count out of the total 33 was greater than 5,000. 30 STUDIES OF KECOlSrSTEUCTED MILK. A bacteriological study of the manufacturing process brought out some useful information. Table VIII gives the number of bacteria pres- ent in reconstructed skimmed milk. These figures, of course, include both the number of bacteria in the water and in the milk powder. Table IV. — Bacterial content of finished product in container. Bac- Bac- Bac- Bac- Bac- Date. teria Date. teria Date. teria Date. teria. Date. teria per c.c. per c.c. per c.c. per c. c. per c.c. 1918. 1918. 1918. 1918. 1919. Sept. 15 7,000 Nov 4 1,200 Nov. 25 1,300 Dec. 13 2,900 Jan. 7 1,200 16 8,000 5 6, 900 26 1,400 2,100 8 1,500 6,000 7,000 200 800 700 18 7,700 6,200 1,300 7,700 10 3, 400 19 10, 000 3,500 500 14 2,000 650 20 1 , 400 6 13, 500 200 16 3,600 23, 000 5, 100 3,700 27 2, 500 28, 000 11 20, 000 2,200 15, 800 200 6,000 21,000 21 3,400 7 5,000 28 800 17 3,300 21,000 23 1,200 2, 000 200 18 3.500 9,600 300 2,500 200 600 4, 600 100 3,100 200 700 5, 500 24 1,000 11 15, 500 29 1,500 19 2,600 14 6,900 25 1,200 12 1,600 700 1 , COO 1,500 26 2,700 3, 200 30 4,000 20 2,200 1,800 30 11,000 8, 000 2,700 2,300 15 17, 500 2,400 1,600 3, 800 21 19,000 3,000 1,100 1,600 Dec. 1 2 , 100 23 15, 600 4,500 Oct. 1 3, 000 3, 000 1,700 4, 100 17 9,400 10 13, 000 13 1,400 500 4,700 1 1 , 000 400 3, 000 2 4,600 24 8,600 12,000 11 13, 500 2,800 500 2,800 18 19, 000 12 7,000 16 4,700 3 15, 000 26 10,000 IS, 000 14 30, 100 6,200 2,000 2,200 55,000 16 1,200 2,700 2,000 700 21 1,600 17 3, 100 2,100 1,000 27 4, 000 200 1,300 1 , 600 1,000 2,000 22 1,700 21 6,000 17 20, 000 600 500 400 4,400 1,100 4 300 28 26, 000 11,000 22 1,200 18 4,500 5 2,000 1,800 23 4,000 5,900 2,200 1,300 900 1,700 2, 100 19 2,400 6 1,600 30 4,400 1,300 23 13, 000 1,100 1 , 000 2,000 26 5, 700 22,000 600 7 3,100 5,500 1,500 24 1,400 2,400 1,500 31 4, 700 3, 100 2,100 200 5, 000 1,200 27 300 28 3, 300 20 2,700 9 1,500 600 1,400 2,800 1,700 1,!-00 600 800 29 12,000 21 600 1,200 1919. 28 12,000 6,800 600 10 4,300 Jan. 2 3,600 700 30 3,000 2,300 1,200 1,300 1,000 500 2,800 800 26, 000 30 1,300 31 2,000 22 1,100 11 3,700 3 1,600 1,000 600 250 3, 100 1,900 700 Nov. 2 2,000 600 1,000 4 2,000 1,000 3 1,000 6, 000 12 5, 000 3,000 31 2,800 4 30, 000 24 1,600 2,900 6 2,300 1,000 3,500 12,000 1,800 3, 200 800 1,000 25 1,900 2,800 500 1,300 1,700 1,300 13 1,200 7 2, 100 Table V. -Results of arranging bacteriological counts of bottled milk according lo magnitude. Range of counts. Number of counts within range. Percent of counts witiiin range. Range of counts. Number ofcounts within range. Per cent ofcounts witliin range. Under 1,000. 1,000-2,500.. 2,500-5,000.. 6,000-7,500.. 7,500-10,000. 10,000-15,000 15,000-20,000 17.4 36.6 21.8 8.1 2.8 5.3 3.2 20,000-25,000 25,000-30,000 30,000^0,000 40,000-50,000 Over 50,000. Total. STUDIES OF RECONSTRUCTED MILK. 31 Table VI. — Bacteriological examinations of reconstructed milk made by the Sanitary Division, Nitro, W. Va. Date. Source. Bacteria per c. c. Date. Source. Bacteria perc.c. 1918. Sept. 23 McssHallNo. 2 2,000 1,650 2,350 2, 150 2,000 1,950 8, 600 1,950 2,600 2,900 2,700 2, 850 900 2, 500 1,.500 1,400 700 1,800 1918. Dec. 16 17 19 20 21 2:^ 24 26 27 28 31 1919. Jan. 1 2 3 4 Hospital mess No 2 200 24 Hospital mess do 5S0 25 do do 1 550 26 do do . 27 do do 1 900 28 do do 2 1.50 30 do do . Oct. 1 do do 2 100 2 do .. ..do 2,450 2,000 2 800 3 do do 4 do do 5 do do Nov. 11 do 12 do 7.50 13 do do 1 1.50 14 do .. ..do ^ . 1 550 15 do do . . 1 200 16 do Table VII.- -ResuUs of rearranging the bacteriological analyses made on the bottled milk by the Sanitary Division, Nitro, W. Va. Range ofcounts. Number ofcounts within range. Per cent ofcounts within range. Underl.OOO 5 27 1 15 1,000-5,000 82 Over5,000 - - -. 3 _._ Total 33 100 Table VIII. — Bacterial content of reconstructed skimmed milk. Date. Bacteria per cc. Date. Bacteria percc. 1918. Oct. 14 17,400 61,000 113, (KK) 27, 600 4,200 7,S00 5,700 5,300 15,000 6, 700 9,000 24,400 41,200 61,010 43, .^00 27, 400 39,000 22,000 1918. Dec. 14 58 000 16 16 246,000 17 19 40 000 22 26 12,100 30 27 23 000 Nov. 4 30 32 000 12 31 8,000 19 1919. Jan. 2 22 25 50,000 26 3 51,000 29 4 15,300 30 11 4,000 Dec. 1 15 126,000 40,000 3 17 9 . IS 280,000 12 23 12,000 13 30 22,000 Rearrangement of these figures, as in Table IX, shows that 23.4 per cent of the counts are below 10,000, and 76.4 per cent are below 50,000. 82 STUDIES OF KECONSTRUGTED MILK. Table IX. — Result of arranging the bacterial counts of samples of reconstructed shimmed milk according to magnitude. Range of values. Number of counts within range. Under 10,000. 10,000-50,000. 50,000-100,000 Over 100,000. Total.. Per cent of counts within range. 8 23.4 18 53.0 4 11. S 4 11.8 100.0 Table X gives the bacterial content of the butter used. A weighed amount of butter was added to sterile dilution water and the mixture warmed until the butter melted. This mixture was then shaken vigorously and the water layer analyzed. The results obtained are somewhat lower than might be expected. Table X. — Bacterial content of butter. Date. Bacteria per c.c. 1918. Oct. 22 38,000 Nov. 4 13,000 22 2,000,000 30 250,000 Dec. 3 280, 000 1919. Jan. 4 120,000 17 650,000 30 35,000 Counts were also made from the milk as it passes from the mixing and pasteurizing vats to the emulsors. These are shown in Table XI. Table XI. — Skimmed milk and butter mixture after pasteurization. Date. Bacteria per c. c. Date. Bacteria per c.c. Date. Bacteria per c.c. Date. Bacteria per c.c. Date. Bacteria per c. c. 1918 1918 1918 1919 1919 Oct. 16 1,600 Nov. 30 400 Dec. 13 520 Jan. 2 3,560 Jan. 22 30,000 17 500 Dec. 2 760 14 660 2 720 22 300 22 1,000 3 4,100 16 800 3 1, 150 23 600 30 500 3 1,100 16 440 4 910 23 890 31 700 3 1,400 18 1,200 6 3,000 26 600 Nov. 4 3,100 4 1,100 18 200 7 3,200 26 780 12 2,600 4 180 19 910 8 600 27 500 12 4,400 7 5,100 21 6,500 10 2,600 27 1,100 19 100 7 600 23 6,000 11 12,600 28 1,800 22 150 9 370 26 100 11 7,100 28 910 22 70 10 24,600 26 580 14 7,900 30 1,400 25 830 10 800 27 100 14 1,400 30 740 25 340 11 14, 000 27 930 15 21,200 31 450 26 310 11 600 28 1,750 15 1,600 26 210 12 3,500 28 305 17 11,000 28 170 12 950 30 820 17 9,400 29 400 12 1,900 31 200 18 40,000 30 1,700 13 2,400 31 580 18 9,300 STUDIES OF RECONSTRUCTED MILK. 33 The killing of the bacteria in the butter and skimmed milk mix- ture depends, of course, on the length of time this mixture is heated. Samples taken at the first stages of a run or immediately after the 30-minute holding period, showed counts slightly higher than sam- ples taken at the end of the emulsification period. This was due, of course, to the fact that the last of the mixture to flow from the mixing vats into the emulsors had approximately 45 minutes more heating at 146° F. than did the first of the batch. Typical results showing the variation in bacteria at the beginning and end of a batch are given in Table XII. Table XII. — Difference in bacterial count at beginning and end of batch from emulsors. Bacteria per c.c. Date. Begin- End. ning. 1918. Dec. 1 1,400 4.-,0 3 2,600 420 11 7,200 720 12 6,000 710 16 1,300 450 26 1,200 780 1919. Jan. 2 1,100 920 11 15,000 3,700 Analyses were also made of the discharge from the emulsors and from the cooler. These figures show little if any variation from the corresponding counts on the milk coming directly from the mixing vats, and are not given. Bacteriology of Cream. Bacteriological counts were also made on the reconstructed cream which was produced at Nitro. A summary of these results is given in Table XIII. These figures, rearranged in the order of their magnitude (Table XIV) show that 74.2 per cent of the counts were below 10,000. Table XIII. — Bacterial content of cream. Date. Bacteria per c.c. Date. Bacteria per c. c. 1918. Nov. 4 2,200 7,000 4,000 1918. Dec. 13 7,700 12 13 3,400 18 17 3,100 18 4,. "00 23 6,600 21 1,500 9,600 400 24,200 25,000 4,300 2,000 12,000 6,000 5,900 5,100 25,000 72,000 27 4,500 23 27 38,000 24 1919. Jan. 2 25 30 5,300 Dec. 2 4 6,000 2 10 6,200 3 13 5,100 7 20 27,000 7 24 50,000 7 25 3,200 10 29 2,100 12 34 STUDIES OF RECONSTRUCTED MILK. Table XIV. — Results of arrangement of bacterial counts on reconstructed cream in order of magnitude Range of values. Under 5,000. 5,000-10,000.. 10,000-30,000. 30,000-50,000, Over 50,000.. Total. Nirmber of counts falling within Per cent of counts falling within range. 12 38.7 11 35.5 5 16.1 2 6.5 1 3.2 100.0 Bacterial Content of Ice- Cream Mix and Ice Cream. There has been a decided controversy recently concerning the bacterial content of well-made ice cream, and many have taken the stand that it is impossible to produce good ice cream with a low bacterial count. At Nitro, ice cream was manufactured in a rather cnide manner, one which would tend to give the product a much higher bacterial content than that produced with modern up-to-date machinery. The bacterial content of the ice-cream produced at Nitro was, therefore, determined with many misgivings, for it was thought that the figures might be misleading. Samples were taken of the mix shortly after it was manufactured and before any storage period had elapsed. Between the manufacture of the mix and the manufacture of the finished ice cream the mix was allowed to stand at a temperature of approximately 33° F. for 24 hours. On manu- facturing, the mix gave 95-100 per cent "swell," showing that it was sufficiently aged to be commercially useful. Table XV gives the bacterial content of the ice-cream mix, and Table XVI that of the finished ice cream: Table XV. — Bacterial content of ice-cream mix. Date. Bacteria per c. e. Date. Bacteria per c. c. Date. Bacteria per c. c. 1918. 1918. 1918. Oct. 12 1,700 Dec. 3 2,200 Dec. 30 4,600 Nov. 5 7,200 3 11,400 11 2,100 7 1.5,000 1019. 16 4,000 9 3,400 Jan. 3 5,000 18 1,300 10 4,600 8 5,100 20 2,200 12 6,400 13 6,700 21 800 14 15,400 17 31,000 26 2,100 18 1,300 20 5,000 26 7,000 21 1 , 700 21 4,000 30 2, SOO 24 6, 500 24 6,000 30 1,600 24 22,000 29 70, 000 Dec. 2 300 28 1,650 30 1,200 STUDIES OF RECONSTRUCTED MILK. Table XVI. — Bacterial content of ice cream. 35 Date. Bacteria Date. Bacteria Date Bacteria per c. c. perc.c. perc.c. 1918. 1918. 1919. Nov. 18 30,200 Dec. 10 37,000 Jan. 20 84,000 21 2,000 12 7, 000 21 9,000 26 10,5, 000 18 18,,SOO 30 23,000 21 27, 200 Dec. 2 16, 000 24 3,000 7 4,500 28 1,800 Rearrangement of these figures shows that 79.5 per cent of the counts on the ice-cream mix were below 10,000, and 71.5 per cent of the counts on the finished ice cream were below 30,000. These figures indicate that, under average conditions, an ice cream can be manufactured with a bacterial content which approximates that of well-pasteurized creams. Butter-Fat Content of Reconstructed Milk. In the operation of a plant manufacturing reconstructed milk, one of the important features is the proper mixing of the ingredients so that the percentage composition of the various constituents are as stated on the final package. The butter used in the making of the products at Nitro was bought with the moisture content guaranteed to be not over 15 per cent. Enough samples of this butter were analyzed to show that the specifications were being complied with, but no detailed analyses were made on the various tubs of butter used from day to day. It was assumed always that the butter con- tained at least 85 per cent fat. Butter-fat determinations made on the finished product brought out the fact that the butter purchased in many cases had a moisture content much less than 15 per cent, making the fat content of the milk greater than was guaranteed. This fact points out the great need for butter analysis, both by the manufacturer and by the purchaser. The moisture content of a sample from every churning, at least, should be determined, and if possible this figure should be stamped on the tubs shipped. In all cases a sufficient amount of butter was added to the milk mix to give 3^ per cent fat, assuming the butter contained 85 per cent butter fat. Analyses of bottled milk were taken a>t first which showed that some variation existed in the fat content of the various containers as put on the market, A study was made to determine whether the fat content of the reconstructed milk at different points in the emul- sification of a single batch varied. The results of these analyses are compiled in Table XVII. Those values which are grouped around definite points have been averaged, giving the values shown in 36 STUDIES OF RECONSTRUCTED MILK. Table XVIII. These values have been plotted in Figure 8. It is easily seen that the fat content at the start is generally less than 3^ per cent, whereas at the end of the run it is much less than 3^ per cent. After 70 per cent of the batch has been run through the emulso]-, there is a decided failing off in the percentage of butfter fat 4.0 3.5 3.0 k b2.5 2.0 1.5 2 OF Batch Discharged Previous to Samplimg. q " 20 40 60 so 2 ^ *^ rf' ^ «^ W*" s A ^ r^ s y y v ■ \ \ \ ^ > L \ > Fig. 8.— Variation in butter-fat content of reconstructed milk at various points in tlie manufacture of a single batch. Nitre, W. Va. in the product. As the vat agitator is in continual motion during the whole process of the emulsion, the reason for the very consistent variation in fat content is not obvious. However, as the condition exists, a cold milk storage vat should be supplied with every installa- tion to allow the equalization of the fat content before bottling. STUDIES OF RECONSTRUCTED MILK. 37 Table XVII. — Results of butter determinations made at various points during the emulsification of a batch of reconstructed milk. Date. Per cent of run discharged previous to sampling. 3 12 13 14 16 25 27 28 29 38 40 42 43 50 53 1918. Dec. 11 3.5 3.4 3.4 3.4 3.1 3.5 3.C. 3.6 14 3.5 3.7 3.7 14 3.7 3.8 3.8 1919 Feb. 8 3.4 3.6 3.7 i'o' 3.8 3.6 3.8 3.6 10 3.8 11 3.8 Date. Per cent of run discharged previous to sampling. 55 57 62 67 68 72 75 80 81 83 86 87 88 93 94 97 100 1918. Dec. 11. ... 3.7 3.6 3.3 2.7 14 3.6 3.3 2.8 14.. 3.9 3.9 3.5 3.2 1919. Feb. 8 . .. 3.5 3.8 3.7 3.6 3.0 2.8 2.3 10 3.7 3.7 3.5 3.7 3.2 3.3 3.0 11 3.8 2.0 Table XVIII. — Average values of butter-fat determinations made at various points during the emulsification of a batch of reconstructed milk. Per cent of run dis- charged Per cent butter fat. sampling. 3.4 14 3.6 27 3.7 41 3.7 54 3.7 66 3.7 74 3.7 84 3.5 95 3.0 100 2.7 Summary. 1. Studies of the bacterial content of this plant's output of recon- structed milk and cream show that it was satisfactory from a health standpoint. 2. It has been demonstrated that satisfactory ice cream can be manufactured on a commercial scale, with a bacterial content com- parable to that of well-pasteurized milk and cream. 3. Using the revolving coil type of ice-cream batch mixer for mix- ing the ingredients entering into reconstructed milk, a product of homo- geneous fat content can not be made, and the final product from an entire batch should be mixed before bottling. o